N3DFix: an Algorithm for Automatic Removal of Swelling Artifacts in Neuronal Reconstructions

Area of Science:

• Neuroscience
• Computational Biology
• Bioinformatics

Background:

• Neuronal morphology is crucial for function, requiring accurate 3D reconstructions from histological data.
• Histological processing can introduce artifacts like swelling, distorting reconstructions and affecting analysis.
• Existing reconstruction methods may not adequately correct these artifacts, impacting morphometric and simulation results.

Purpose of the Study:

• To introduce N3DFix, an open-source software for automatically detecting and correcting swelling artifacts in 3D neuronal reconstructions.
• To provide a post-processing tool compatible with both manual and semi-automatic reconstruction workflows.
• To validate the performance of N3DFix against manual correction methods.

Main Methods:

• Developed an open-source software, N3DFix, employing a correction algorithm for swelling artifacts.
• Trained the algorithm using a "ground truth" dataset created by neuroanatomists performing manual reconstructions with and without artifact correction.
• Validated N3DFix's performance by comparing its automated corrections to meticulous manual corrections.

Main Results:

• N3DFix successfully identifies and corrects swelling artifacts in 3D neuronal reconstructions.
• The software's performance in artifact correction is comparable to that of expert manual correction.
• N3DFix is available as a post-processing tool for NEURON, Vaa3D, and Py3DN platforms.

Conclusions:

• N3DFix offers an effective and automated solution for correcting swelling artifacts in neuronal reconstructions.
• The software enhances the accuracy of morphometric analysis and computational simulations by improving data quality.
• N3DFix promotes wider adoption of accurate neuronal reconstruction techniques through its accessibility and performance.